Legacy Code: Big Rewrite or Progressive Rejuvenation?
summary
Argues against a big rewrite and for progressive rejuvenation on a large finance system, citing three failed rewrite attempts and a recurring level 2 emergency every week. First steps: simplify onboarding, create 20 end-to-end tests that capture current input→output behavior, then split the monolith into modules and apply the strangler pattern. Gives a real migration path (example: C# to Scala + JavaScript), shows incremental feature moves and data migration pitfalls, and reports lower risk and steadier delivery than wholesale rewrites.
AI-written summary · about
chapters (7)
- 0:00Start
- 0:49Legacy Systems
- 2:48What is a legacy system?
- 6:09Technical Debt Default
- 20:08Code quality
- 21:39Tea House Western Architecture
- 22:39Wabi Sabi Architecture
related talks
41mRefactoring your legacy applications the right way
45mChanging attitudes toward legacy code
11mThe monster under the bed - working with legacy code
33mTaylor’s Guide to Big Rewrites
27mBig Rewrite Strikes Again
42mHow to sell a big refactor or rewrite to the business?
13mThis Place is a Mess, Rewrite Everything!
15mBuilding a Modernization Strategy